Windows on ARM and the future of PCs as a Service

The looming battle has implications for the future of Windows and the future of both PCs and smartphones.

thinkstock 500773792 cpu processor

This month, the initial battle will begin on what will be an historic war for the next generation of PCs.  Initially it will be fought on laptops but – much like smartphones drifted to tablets, and much of the initial wave of AI-driven, home-based digital assistants – this war may eventually encompass all PCs.

If this were just a war between processors, the X86 folks would likely win easily, both because Windows has decades of tuning on X86 and it is the entrenched part. But this isn’t about processors. This is about whether the computing will be done on the desktop or in the cloud.

In short this is a war between the modem and the processor…or yet another attempt to turn the PC into more of large smartphone.

This has a lot of implications both for the future of Windows and the future of both PCs and smartphones. Let’s look at some of the interesting moving parts.

PCs as a Service

One of the big trends this decade is PCs as a Service – and every major OEM is racing to implement their own version. The idea is much like it was with mainframes in the early days (and used to be with cell phones while subsidies were the norm). You don’t pay for the hardware, it’s built into a monthly charge, which includes the user’s apps and the hardware needed to run them. The end goal is to have something financially that behaves far more like a cable service or utility than a typical PC sale.

The problem has always been the hardware, and the concern that, after getting new hardware and paying for one or two months, the company cancels the service, leaving the OEM with old hardware they must sell for pennies on the dollar. In addition, if the service is to really work, the cost of operating that hardware must be low, and more in line with a set top box or terminal than a typical PC.  

So, the implementation would tend to favor hardware that was more terminal- or thin client-based, so the complexity as well as the OEM’s service costs could be reduced, and thus the fee they would have to charge, minimized.

The cloud

The gating factor for these future PCs as a service product would likely be the connectivity of the device and less about the processor.  And clearly one of the other big trends that helps us down this path is the spread of cloud computing and solutions like Citrix, which could be used to supply the needed user experience. Given Microsoft’s move to make Azure their most important growth platform, this idea of cloud desktop computing should be core to their long-term strategy anyway. It would also solve at least one huge problem they currently have. People not patching their machines timely or upgrading to the current version of their OS platform as both would be centrally controlled and implanted. 

But we’ll need better wireless performance.


This performance is coming in 5G before the end of the decade from almost every carrier. And 5G not only brings more bandwidth, it provides better connectivity at the edge – which means you should be able to get decent performance regardless of where you are. One other benefit is security, because we’ve there are a lot of compromised WiFi access points, and being able to connect directly back to a cellular provider should significantly improve security over laptops that currently use WiFi. There are a few other advantages in that the result should be far more like a smartphone: always-on so you don’t have to wait for it to boot or to come out of sleep or hibernation. 

Wrapping up: thinking different

Running against entrenched vendors with decades in a segment isn’t easy. Apple did it with the iPhone by fundamentally changing what an iPhone was. Prior to the iPhone, smartphones were business-centric devices and bought by companies. After the iPhone, they were consumer-centric and increasingly bought by users (even though, in many cases, companies still reimbursed the purchase).

This is what the ARM version of Windows will have to do. In this case, shifting buyers from largely computing locally to cloud computing, and from mostly being connected to always being connected. 

The one gating factor I haven’t addressed is airplane use. Right now, the truly crappy connections available on airplanes just isn’t enough to provide for a cloud computing desktop model. This capability is scheduled to be significantly improved, but given how bad it is now, I have doubts whether these improvements will be enough. Given how tight the physical space on airplanes have become, it isn’t clear many of us will be able to truly get real work done on airplanes much longer anyway, so this may turn out not to be a critical path problem. 

In the end, we are moving to new model in personal computing. That new model puts existing hardware at risk, and provides an unusual opportunity for ARM technology to displace x86.

Now it is just up to someone to do the Steve Jobs thing and execute.

If ARM can get a beachhead on laptops, then desktops, which are generally always connected, should fall fast. Then, if the PC becomes a bigger, connected smartphone-like device, will we still also need smartphones, or will we simply have a smaller accessory connected to our connected PC that we talk through? I wonder…

[Disclosure: Microsoft is a client of the author.]

Copyright © 2017 IDG Communications, Inc.

8 simple ways to clean data with Excel
Shop Tech Products at Amazon